Duct Noise Research Articles

Suppression of broadband noise radiated by a low-speed fan in a duct

Attenuation of ducted fan noise remains a technical challenge in the low frequency range as traditional duct lining becomes ineffective. This study proposes a reactive method to suppress the sound radiation from an axial-flow fan. The method is particularly effective in the low frequency region and covers a broad band. Its effect is derived from two mechanisms. One is the reduction in the confining effects of duct walls when the duct radius is increased; the other is the acoustic interference between the direct radiation from the fan and reflections by the duct junctions. This interference is always destructive for axial dipoles when the frequency approaches zero. This performance differs from normal passive control methods, which become totally ineffective toward zero frequency. An approximate plane-wave theory explains the essential physics of the method, and its quantitative prediction is found to agree well with a full numerical simulation using a spectral method of Chebyshev collocation. The latter is validated by experiment using an axial-flow fan in a duct of finite length. Broadband noise reduction is achieved while the flow speed is kept unchanged. Practical difficulties of implementation for a fan with high pressure increase are discussed.

106 複数の近接した周波数成分が含まれるダクト内騒音の能動制御(騒音・振動の能動制御(2))

The present paper describes the experimental investigation of characteristics of adaptive feedforward active control systems of beat noise in ducts. Dual-input, dual-controller and single-output(2-2-1) system has been proposed to improve the performance of the active control. Proposed 2-2-1 system can achieve relatively large attenuation and can converge rapidly in comparison with the conventional single-input, single-controller and single-output system. Results show that the step-size of the 2-2-1 system increases as the frequency separation increases and noise reduction increases as the step-size increases. Results show that the fastest convergence is obtained by the optimum step-size for a given frequency separation and sampling frequency.

Multiple drumlike silencer for low frequency duct noise reflection

Low frequency duct noise is difficult to tackle by passive means within a certain space limitation. Drumlike silencer offers broadband sound reflection in the low-frequency region, e.g. 0.1–0.3 times the first cut-on frequency of the central duct. Optimization was carried out for the logarithmic bandwidth in which the transmission loss (TL) is greater than a pre-determined criterion value. The prediction was validated in laboratory but it was anticipated that the required high level of tension is difficult to implement in practice. It is found that the optimal tension is approximately proportional to the 3rd power of membrane length, so a partitioned membrane is expected to reduce the required tension. In order to optimize the partitioned drumlike silencer, a new cost function of total transmission loss, or insertion loss, with a frequency weighting in favour of a certain wide band of low frequencies is introduced. The new cost function removes the undesirable discontinuous behaviour occurring in the previous optimization process. It is also found that the partitioned silencer can drastically improve the weighted total TL and the tension required is indeed much reduced. The main performance improvement derives from around the lower frequency limit of interest.

Attenuation of low frequency duct noise by a flute-like silencer

A broadband, duct noise reflection mechanism is introduced in this theoretical study. It consists of side-branch cavities filled with a light gas, e.g. helium, and covered by impervious, tensioned membranes as two apertures, one at the inlet and another at the exit. Incident waves are scattered by the membranes into two passages, one through the central duct and another through the cavity bypass. Due to the faster speed of sound in the bypass, a Herschel–Quincke tube resonance appears and gives a peak in the transmission loss spectrum. Another resonance occurs when the frequency of the incident sound coincides with the vibroacoustic frequency determined by the membrane tension and inertia contributions from the membrane and the fluid media. With appropriate tensile stress, the trough between the two spectral peaks can be elevated to a desirable high level, e.g. 10 dB, and the crucial factor is identified as the low density of the cavity gas filling. The broadband sound reflection performance is comparable with and even exceeds that of the drum-like silencer [L. Huang, Parametric study of a drum-like silencer, Journal of Sound and Vibration 269 (2004) 467–488] with the same cavity geometry, but the current mechanism requires a low tensile stress which is much easier to implement in practice.

The prediction of airflow generated noise in ventilation system bends.

This paper describes recent work on the prediction of airflow generated noise in ducts based upon pressure loss characteristics. Strategies employed to reduce turbulence by smoothing the flow around a bend include the use of turning vanes, and a large radius of curvature with the associated use of splitters. The effect of these strategies on the regenerated noise is examined for a range of common bends including mitred bends, with and without turning vanes, and radiussed bends, with and without splitters. The effect of bends with different aspect ratios is also examined. An excellent collapse of measured data onto a single characteristic curve is reported for each configuration based upon standard values of pressure loss coefficients. It is suggested that that these characteristic curves might form the basis for a practical prediction method.

하이브리드 제어 알고리즘을 이용한 덕트내 능동소음제어

This study presents the active noise control of duct noise. The duct was excited by a steady-state harmonic and white noise force and the control was performed by one control speaker attached to surface of the duct. An adaptive controller based on filtered x LMS(FXLMS) algorithm was used and controller was defined by minimizing the square of the response of the error microphone. The assemble controller, which is called a hybrid ANC(active noise control) system, was combined with feedforward and feedback controller. The feedforward ANC attenuates primary noise that is correlated with the reference signal, while the feedback ANC cancels the narrowband components of the primary noise that are not observed by the reference sensor. Furthermore, in many ANC applications, the periodic components of noise are the most intense and the feedback ANC system has the effect of reducing the spectral peaks of the primary noise, thus easing the burden of the feedforward ANC filter.

A Module-Based Active Noise Control System for Ventilation Systems, Part II: Performance Evaluation

To utilize the full noise attenuation potential of an active noise control (ANC) system applied to duct noise, it is important to be able to minimize the turbulence induced noise in the microphone signals. This is the second paper in a series of two, that treats the problem of turbulence induced noise originating from the airflow inside the ducts, when applying ANC to ducts. Part I contains theoretical and experimental investigations of the influence of the turbulence induced noise on the filtered-x LMS algorithm used in the ANC system. Part II (the present paper) is concerned with design and investigations of microphone installations which produces a sufficient amount of turbulence suppression while also meeting industry requirements. These requirements are, for example, that the microphone installations should be based on standard ventilation parts, and that they should be easily installed and maintained. Furthermore, results concerning the performance of an ANC system with different microphone installations are presented. Some of the results were obtained at an acoustic laboratory according to an ISO standard. The attenuation achieved with ANC was approximately 15-25 dB between 50-315 Hz even for airflow speeds up to 20 m/s.

Broadband duct noise attenuation by a middle-ear mechanism.

Low-frequency duct noise is difficult to control passively by any device limited by its total volume due to system stiffness. When a side-branch device is used, cavity air becomes incompressible toward the dc frequency. One way to tackle the relative incompressibility for low to medium frequencies is to introduce acoustic phase difference in the cavity air oscillation, which otherwise behaves like the low-order room acoustics modes. The introduction of a light plate covering the cavity in our previous studies [J. Acoust. Soc. Am. 119, 2628–2638 (2006)] does essentially that by enhancing the non-volume-displacing modes. Optimization shows that the required stiffness-to-mass ratio is much beyond what is available for existing bulk materials. In the present study, an alternative device that resembles the ossicular structure of the mammalian middle ear is introduced. The acoustic signal is passed along through a rigid pin which simulates the ossicular bones via its longitudinal motion. Contrasting with the usual flexural waves in plate or membrane structures, the bone transmission has almost infinite speed and it creates the phase difference needed to overcome the system stiffness. Theoretical predictions show that the performance of such device is comparable and better than the idealistic plate silencer with interesting comparisons of features.

A numerical study on multimode sound propagation in lined ducts and radiation to the far field

In previous articles, the authors developed a hybrid scheme for analysing bypass duct noise, in which a numerical analysis using finite element method for in-duct propagation and an analytic radiation code with fully represented effects of bypass shear layer are coupled. Such procedure permits detailed study on the interaction between duct configurations, such as geometry and acoustic liner impedances, and modal propagation and attenuation, and also the effects on the radiation pattern, within practical timescale and at modest computational cost. The scheme has been applied to realistic aero-engine bypass ducts and has been integrated with an optimisation programme. The numerical results obtained so far have revealed that for ducts with acoustic liners highly attenuated modes are not necessarily those with high mode angles, which is contrary to general anticipation. The aim of the study in the current paper is to understand the physics behind this phenomenon and its effect on the radiation to the far field. Detailed investigation on modal attenuation is performed by using the hybrid scheme. The effect of modal power distribution at the duct exit on the directivity pattern of the radiated noise to the far field is also discussed.

Low Frequency Acoustic Performance in Ducts Using Honeycomb Layer Systems

The use of composite materials, like honeycomb structures, in the building industry is growing rapidly, due to their superior performance (very high Sound Transmission Loss) and significant weight reduction over conventional materials. An experimental investigation into low-frequency duct noise reduction using honeycomb sandwich panels is presented. Composite sandwich structures with an air gap between panels provide low-frequency sound absorption. The waves in the duct are coupled with those in the porous layer and the air cavities. The energy is dissipated as a result of the resonant mass-spring behavior of the system. A significant impact on the noise reduction is obtained by using a comparatively stiff honeycomb panel, in combination with a thin metallic duct wall with minimum weight. A noise reduction up to 12dB, below 300 Hz, is demonstrated.

Noise Generation in Ventilation Systems by the Interaction of Airflow with Duct Discontinuities: Part 1 Bends

Discontinuities in air duct systems give rise to turbulence and some of this turbulence results in airflow generated noise. One potential source of flow generated noise in ducts is the presence of a bend. Various strategies are employed to reduce turbulence by smoothing the flow around a bend, such as the use of turning vanes, large radius of curvature and the associated use of splitters, and these will also affect the noise generated. Recent work on the prediction of airflow generated noise in ducts based upon pressure loss characteristics is applied to a range of bends including mitred bends, with and without turning vanes, and radiussed bends, with and without splitters. The effect of bends with different aspect ratios is also examined. An excellent collapse of measured data onto a single characteristic curve is reported for each configuration based upon standard values of pressure loss coefficients. It is suggested that these characteristic curves might form the basis for a practical prediction method.

Effects of Cross-Sectional Dimensions on Active Noise Control in Rectangular and Round Ducts

Active noise control (ANC) works best to reduce low frequency noise. Because many industrial noise sources are broadband, ANC may be used more if it can be successfully applied to higher frequency ranges. This study explored one method to increase ANC effectiveness at higher frequencies. ANC is particularly useful in hard-walled ducts where plane waves propagate. Higher order mode waves are much more difficult to control. Basic acoustic principles dictate that the cut-on frequency at which higher order modes will first begin to eclipse simple plane waves in a duct will be determined by the cross-sectional geometry of the duct. The lowest frequency for higher order modes increases as duct diameter decreases; therefore the range of frequencies where plane waves dominate will be greater and effective control using ANC will be better as duct diameter decreases. The result is that somewhat higher frequencies can be controlled with ANC for smaller diameters. Below the first higher order mode cut-on frequency for the largest size studied, there should be little difference in ANC effectiveness between the duct sizes. To test those suppositions, a commercially available ANC system was used to reduce random noise in rectangular and round ducts having different diameters. Results showed that insertion loss (IL) ranged from 5 dB to 29 dB in frequencies ranging from 40–1000 Hz and varied inversely with cross-sectional size as expected. There was no difference in IL below 280 Hz (p = 0.7751) between the different diameter ducts. There was a significant difference between duct diameters above 280 Hz (p < 0.0001). The same tests were conducted on a rectangular duct with one cross-sectional dimension fixed and one varied at seven different sizes. Results showed similar IL from 5 dB to 29 dB that varied inversely with size.

Modelling and control of noise in a flexible flow duct

Stability and control of noise in a flow duct are discussed in this paper. The flow is assumed to be uniform for the steady state. But an unstationary interaction between acoustic waves and a flexible part of the wall can induce wave amplification up to an instability. This phenomenon is similar to flutter encountered in aeroelasticity of airfoils [H. Dowell, C. Curtiss Jr., R.H. Scanlan, F. Sisto, A Modern Course in Aeroelasticity, Kluwer Academic Publishers, 1989; Y.C. Fung, An Introduction to the Theory of Aeroelasticity, J. Wiley, New York, 1955]. In fact, one can suggest that the flexible wall can act as a trap for acoustic energy and using a smart device located on this structure, one can design an efficient control method for noise. The goal of this paper is to describe a simple model and few numerical tests which partially justify these considerations.

A moving zone of quiet for narrowband noise in a one-dimensional duct using virtual sensing

A frequent problem in active noise control is that the zone of quiet created at the error sensor tends to be very small. This means that the error sensor generally needs to be located close to an observer's ear, which might not always be a convenient or feasible solution. Virtual sensing is a method that can move the zone of quiet away from the error sensor to a desired location that is spatially fixed. This method has been investigated previously, and has shown potential to improve the performance of an active noise control system. However, it is very likely that the desired location of the zone of quiet is not spatially fixed. An active noise control system incorporating a virtual sensing method thus has to be able to create a moving zone of quiet that tracks the observer's ears. This paper presents a method for creating a moving zone of quiet based on the LMS virtual microphone technique. To illustrate the proposed method, it is implemented in an acoustic duct and narrowband control results are presented. These results show that a moving zone of quiet was effectively created inside the duct for narrowband noise.

Active Helmholtz Resonator With Positive Real Impedance

The impedance of a passive noise control device is strictly positive real, if the device is installed in noise fields with weak mean flows. Passive noise control devices are, therefore, more reliable than active ones. Active control may be applied to a Helmholtz resonator to introduce electronic resonance. It will affect the impedance Zact of the resonator. A controller may be designed such that (a) Zact is small and resistive at some tunable frequencies; and (b) Re{Zact}⩾0 in the entire frequency range of interest. If criterion (a) is satisfied, the active resonator can suppress duct noise at tunable frequencies. It is difficult to design a controller to satisfy criterion (b) because parameters of the controller depend on acoustic parameters of the noise field. A new method is proposed here to design an active controller to meet both criteria simultaneously. The satisfaction of criterion (b) implies a positive real Zact and a robust active resonator with respect to parameter variation in the noise field. Experimental results are presented to verify the performance of the active resonator.

Critical assessment of thermal/acoustical insulation for pipe and duct noise

Piping and duct networks are among the major sources and distributors of noise in power plants. This noise is generated primarily by large rotating equipment and devices, primarily valves, with a large pressure drop. Piping and duct noise controls include silencers, thermal/acoustical insulation (lagging), and special treatments (e.g., low noise valve trims and diffusers). The effectiveness of acoustical lagging for pipe and duct noise depends on the type of source, coupling of the piping to sources or support structures, and the frequency spectra radiated. The acoustical performance of pipe lagging systems has been reported by a number of researchers over the past several decades. Today both ASTM and ISO have issued testing standards for the evaluation of pipe lagging systems, and the ISO standard also provides pipe lagging design guidance. Current technology for measuring the TL of lagging, predicting pipe and duct noise, and predicting the IL of lagging is critically assessed. This paper reviews the methods and results of the two standards and compares several laboratory test results with field measurements. Some challenges are presented along with suggestions for improving prediction of piping and ductwork noise abatement within power plants and in the community.

“Buzz-saw” noise: A comparison of measurement with prediction

Accurate prediction of “buzz-saw” noise in a turbofan inlet duct necessitates consideration of nonlinear acoustics, modelling a complete fan blade set, modelling an acoustic liner, and calculations at high frequencies. A recent series of papers has described new work concerning the application of one-dimensional propagation models to the prediction of buzz-saw noise. A numerical model, termed the frequency domain numerical solution or FDNS, has been developed. It can be used to calculate the nonlinear propagation of the rotor-alone pressure field in either a rigid or acoustically-lined inlet duct. From this the in-duct noise level of the buzz-saw tones can be determined. In previous work, validation of this method by comparison with in-duct noise measurements has been limited to rigid inlet ducts, because of the lack of availability of suitable measurements from lined ducts. In this article new measurements of buzz-saw noise in an acoustically-lined inlet duct are utilized. A comparison of measurements of buzz-saw noise in a lined inlet duct, and noise predictions from numerical simulations by the FDNS is presented. The detailed measurements reveal the effect of an acoustic liner on buzz-saw noise. The suitability of the numerical model to be used to provide realistic noise predictions for supersonic ducted fans is also examined.

A selected history of duct mode synthesis

Attempts to attenuate noise propagating down a duct, with or without flow, must recognize that wall boundary conditions permit only certain spatial wave patterns (modes) to exist. The aircraft noise research community began building mode synthesizers at least thirty years ago in order to study duct propagation and radiation phenomena in a laboratory environment. Lockheed Georgia built a spinning mode synthesizer (SMS) for NASA in the 1970s. NASA used its SMS in a circular flow-duct to validate predictions of mode propagation through constrictions, lined sections, and inlets of various shapes. In the mid-1970s, Penn State created a mode synthesizer as part of a demonstration of active noise control in a circular, no-flow duct. NASA sponsored a series of studies in the 1990s aimed at maturing active control technology for ducted fan noise. Each of these active control systems was essentially a mode synthesizer coupled with a control system to cancel fan-generated noise. NASA is currently building a new mode synthesizer in a rectangular duct to study the effects of high-speed flow and curvature of the duct axis on advanced noise suppression technology. The evolution of mode-generation schemes will be discussed.

Application of H-infinity hybrid active controller for acoustic duct noise cancellation

This paper describes the principle and application of a hybrid robust active noise controller (ANC) for reducing broadband noise in acoustic ducts. The proposed ANC system consists of an adaptive controller with filtered-x least mean square (FXLMS) algorithm and a feedback structure with robust synthesis theory for obtaining robust performance and stabilisation. In an active control structure, plant uncertainty is one of the major factors that may affect the performance as well as stability of a system and must be taken into account in controller design. In this study, the controller is designed by frequency domain method and implemented on a digital signal processor (DSP) platform for real time control. Experiments are carried out to evaluate the attenuation performance of the proposed control system for synthetic random noise in an acoustic duct. The experimental results indicate that the proposed hybrid controller is effective in suppressing the undesired broadband noise. Furthermore, characteristic analysis and experimental comparison of adaptive feedforward structure, pure feedback control structure and proposed hybrid structure for reducing the noise in a duct are also presented in this paper.

Multi-Channel Adaptive Feedforward Control of Noise in an Acoustic Duct

The single-reference/multiple-output active noise control (ANC) of the accurate physical model of an acoustic duct system, developed earlier by the authors [1] is investigated. An adaptive feedforward algorithm is developed that minimizes a generic cost function consisting of the p-norm of the error vector. A computer model of a multi-channel ANC system, with the tonal and sweep sine input signals, is established, and the results for different single-input/single-output (SISO) configurations (error microphone and secondary source locations) of the ANC system are compared. The dynamic response of the single-reference/multiple-output ANC systems, using the Minimax and MEFXLMS algorithms, is studied and the effect of the acoustical feedback neutralization in the multi-channel ANC system is examined. Simulation results demonstrate that the multi-channel adaptive feedforward ANC system with the Minimax algorithm has a superior performance to that of the same system with the MEFXLMS algorithm. However, the total noise reduction of the Minimax algorithm, compared with that of the MEFXLMS, is much dependent on the frequency content of the input signal.